1

A Disciplined Intelligence: Critical Inquiry and Canadian Thought in the Victorian Era McGill-Queen’s University Press A. B. McKillop Year: 2001 Language: english File: PDF, 18.02 MB Your tags: 0 / 0

2

Ethnology of the Ungava District, Hudson Bay Territory McGill-Queens University Press Lucien M. Turner Year: 2001 Language: english File: PDF, 19.39 MB Your tags: 0 / 0

JANUA LINGUARUM
STUDIA MEMORIAE
N I C O L A I VAN W I J K DEDICATA
edenda curat
C. H.

SCHOONEVELD

Indiana University

Series

Practica,

133

LATIN PUNCTUATION
IN THE
CLASSICAL AGE
by

Ε.

OTHA

WINGO

1972
MOUTON
THE HAGUE · PARIS

© Copyright 1972 in The Netherlands.
Mouton & Co. N.V., Publishers, The Hague.
No part of this book may be translated or reproduced in any form, by print, photoprint,
microfilm, or any other means, without written permission from the publishers.

LIBRARY OF CONGRESS CATALOG CARD NUMBER: 79-159474

Printed in Hungary

PARENTIBVS
PIEN TISSIMIS
ET
CARISSIMAE
VXORI

ACKNOWLEDGEMENTS

I wish to express my deep appreciation to my adviser, Professor Revilo P.
Oliver, for suggesting the problem and for his constant and patient guidance
and painstaking assistance throughout the research and writing of the dissertation;
to Professors Ben E. Perry, Alexander Turyn, Antonio Tovar, and Gertrude
Smith, as well as to Professor Oliver, for the challenging classes and encouraging
conferences throughout my graduate work at the University of Illinois;
and to Dr. Edith Carrington Jones, who was always generously helpful
in locating and using materials in the great Classical Library at the University
of Illinois, which was founded and in large part assembled by the late Professor William Abbott Oldfather.

TABLE OF CONTENTS

Acknowledgements
I. The Problem of Punctuation
II. The Testimony of the Grammarians

7
11
20

III. The Res Gestae Augusti

29

IV. Latin Books

50

V. Legal Style

68

VI. Other Styles

83

VII. The Signs of Punctuation

94

VIII. Conclusions

132

Appendix

Papyri Latinae

134

Metrical Punctuation

140

I.

Appendix II.
Bibliography

164

I
THE PROBLEM OF PUNCTUATION

This treatise is an attempt to answer, insofar as it is possible, the question
whether or not the classical writers, such as Cicero or Vergil or Livy, were
read by their contemporaries in copies that had punctuation. I shall give,
therefore, as complete an account as possible of the types and fun; ctions of
punctuation in Latin writing of the classical age. Every published inscription,
papyrus, and manuscript that antedates the fall of the Roman Empire in
the West has been considered, provided that enough of the text remains
to be intelligible and thus show the meaning of punctuation where that is
used. For this study I have checked every inscription in the Corpus Inscriptionum Latinarum and in L'Annee Hpigraphique (1944—1961) and every
papyrus listed in Cavenaile's Corpus Papyrorum Latinarum, using also such
facsimiles as were available in other publications whenever there was reason
to believe that the original contained punctuation.
As late as the year 1908 so careful a scholar as Ioannes Ferrara knew so
little of punctuation in the classical age that he described the marks of punctuation in the fragment of the Carmen de Bello Actiaco as signa "quae quid
spectent me plane ignorare fateor". 1 He thus reached the odd conclusion
that those signs must mark places which the author intended to revise. I t
followed, therefore, that the papyrus was the author's autograph of a poem
that had never been completed. Therefore the poem could not be the work
of any of the ancient poets known to have composed poems on that subject.
Hence it was probably the work of the man who owned the villa in Herculaneum at the time of the eruption in 79 A.D. 2
The manuals of epigraphy and palaeography summarily dismiss the subject
of punctuation with brief mention of the interpunctum, or medial point,
1

Poematis Latini Fragmenta Herculanensia (Papiae, 1908), p. 33.
This scaffolding of inferences has, of course, long since collapsed. The latest editor
of the fragments (Garuti, 1958) returns to the early attribution of the poem to Gaiue
Rabirius and supports the attribution with arguments on which he relies so confidently
that he places the name of Rabirius on his title page (C. Rabirius: Bellum Actiacum).
1

12

THE PROBLEM OF PUNCTUATION

used as a word-divider. For the sake of completeness, I shall review the meager
information given in the standard reference works.
James C. Egbert, Introduction to the Study of Latin Inscriptions (1906),
says that "from earliest times individual words in inscriptions were separated
by marks of punctuation, which regularly occupied a position midway between
the upper and lower limits of the letters. These points, as a rule, were not
placed at the end of a line, and did not indicate sentences or parts of sentences."
After discussing the various forms of the medial point, he remarks that "In
carmina the metric lines are marked, and when half-verses are written this
also is indicated by punctuation. At times a long space takes the place of the
punctuation mark. In one of the Scipio 'Elogia' the ends of the metric lines
are indicated by the spacing, while in another the punctuation mark is used,
in this case a horizontal line."
Rene Cagnat, Cours d'Epigraphie Latine (1914), discusses the shapes of
the "points separatifs", but mentions no use of any mark of punctuation
to show syntactical relation or sense.
John Edwin Sandys, Latin Epigraphy (1927), dismisses the matter of
punctuation by stating that the medial points "are not used to denote the
termination of a clause or a sentence."
Pedro Batlle Huguet, Epigrafia Latina (1946), merely mentions interpuncta as word-dividers, and adds that "son rarisimos los ejemplos con dos
ο mas puntos puestos en linea vertical u horizontal".
Edward Maunde Thompson, Introduction to Greek and Latin Palaeography
(1912), surveys Greek punctuation; his discussion of Latin punctuation is
confined to later centuries. "The punctuation of Latin MSS. followed in some
respects the systems of the Greeks. In the poem on the Battle of Actium,
found at Herculaneum, points are used to mark off the words, a practice
borrowed from inscriptions; and in the early MSS. of Virgil in the Vatican
Library points are found employed for the same purpose, although they
appear to be due to a second, but still early, hand. From the Latin grammarians we know that they adopted the Greek system of punctuation by points
(θέσεις, positurae), to which they gave the titles of 'distinctio finalis', 'subdistinctio', and 'distinctio media', but in practice we find that the scribes
used the points without consistently adhering to their meaning." He does
mention the use of a short space, in some of the more ancient manuscripts,
to indicate the conclusion of a passage or paragraph, and a paragraph mark,
used to separate paragraphs or divisions of the text, as in the poem on the
Battle of Actium.
Giulio Battelli, Lezioni di Paleografia (1949), does not consider the use
of punctuation before the Eighth Century.
Jean Mallon in his Polygraphie Romaine (1952), in which he deliberately
sets out to revolutionize the study of early Latin palaeography, strangely

THE PROBLEM OF PUNCTUATION

13

makes no mention of punctuation as an indication of date or type, and even
ignores marks other than the interpunct in his transcriptions of some of the
examples shown in the plates appended to his work.
E, A. Lowe, Codices Latini Antiquiores (1934 — 1963) sums up the situation
by repeated statements to this effect: The history of punctuation in Latin
manuscripts is a chapter still to be written. He cautions that punctuation
"can be studied only by inspecting the original, and even then it is not always
easy to determine what is due to the first hand and what to later correctors:
facsimiles often give a false impression . . . ," 3
There is no discussion of Latin punctuation in either Müller's Handbuch
or in Pauly-Wissowa-Kroll. Aemilius Hübner, Exempla Scripturae Epigraphicae Latinae (1885), mentions points in carmina; the paragraph signs
7 and 3 in the Monumentum Ancyranum; "virgulae ad commata separanda
inter litteras interpositae" in the Laudatio Murdiae; and spaces for separation
of phrases or paragraphs; as well as various signs placed at the ends of lines.
However, he does not discuss the significance of these marks.
Only recently has the importance of the use of spacing to indicate divisions
of thought and syntax been in any way emphasized—by Joyce S. and Arthur
S. Gordon, Contributions to the Palaeography of Latin Inscriptions (1957).
However, they do not go into great detail, since their concern is with dating
of inscriptions. The Gordons discuss arrangement in paragraph form, and
go into details concerning the use, position, and formation of the interpunctum
as a simple word-divider—and the omission of it. Their chapter headed
"Punctuation" deals almost entirely with the interpunct. They consider
other punctuation only in connection with the occurrences of it on the inscriptions published in their Album of Dated Latin Inscriptions (1958).
These scholars leave us with little definite information on the subject of
punctuation. The impression given by their comments is that Latin writing
did not utilize syntactical or sense punctuation at all.
It would seem that an understanding or appreciation of the use of punctuation in the classical age is not much more widely diffused among scholars
in our day than it was in Ferrara's. Marcel Durry, in what is the latest edition
of the text of the so-called Laudatio Turiae (Paris, 1950, with a sheet of addenda
and corrigenda, c. 1952), prints in II.2a-3a of the text a restoration (his own)
which simply ignores the punctuation that is clearly visible on the stone
before the preceding word. The stone shows that a new clause began with
the word ornamentis, but Durry's restoration makes the new clause begin
after that word, although restorations that would not violate the punctuation

3 L o w e is thinking primarily of the vellum codices in which the occasional points to
set o f f words appear usually, if not invariably, to have been added b y a hand later than
the scribe's; cf. infra, p. 23 note 11.

14

T H E PROBLEM OF

PUNCTUATION

(e.g. by de Sanctis, reported in Durry's apparatus) are possible. I shall not
here attempt to decide what is the correct restoration: I merely point out
that it apparently did not occur to Durry that the punctuation had any significance at all.
The term 'punctuation,' in the restricted sense in which I shall use it here,
refers to the use in writing of certain signs to show the end of a sentence
or to indicate its structure or the interrelation of its parts for the sake of
clarity to facilitate reading. Marks may also be inserted, again for the sake
of clarity, to show pauses in speech even when no syntactical consideration
would demand them. In a somewhat broader sense, a system of punctuation
includes the arrangement of the material into paragraphs by using spacing
and indentation to mark off units of subject matter. And in the broadest
sense, any sign that is not a letter of the alphabet is a mark of punctuation.
I may therefore delimit the topic by listing related matters which fall
outside the scope of this dissertation.
1. The interpunctum as a word-divider. This falls outside our topic because
word-division was universally used during the period in which we are interested
and is therefore to be taken for granted. Indeed, what will be significant for
our purposes will be the absence of interpuncts, which, unless there are contrary indications, will place the text outside the historical period in which
we are interested.
The division of writing into words is a practice of great antiquity, and
appears to have been customary in the writing of various languages in cuneiform characters. The practice may have reached the classical peoples through
Crete, for word-division is clear on the Phaistos-disk, which may be as early
as 1700 B.C.4 The practice of word-division was carried into Greek when the
Cretan characters were used to write that language, for Sterling Dow notes
the presence of the interpunct in Linear B, "regularly inserted, as in cuneiform, between words". 5 I t is possible that the practice of word-division
somehow survived through the Dark Age that followed the fall of Mycenaean
culture and so persisted when the Minoan characters were replaced by the
Greek alphabet. M. Lejeune in B.JS.A.,
LVI (1954), p. 429, remarks: "les
plus anciennes inscriptions grecques connaissaient l'usage des interponctions
sdparant mots ou groupes de mots, et il est remarquable que les Grecs aient
4

The Phaistos-disk must antedate the oldest specimens of Linear Β (in which an
interpunct replaces the vertical line as word-divider). Leonard R. Palmer, Myceneans and
Minoans (London, 1961) dates the beginning of Linear Β to c. 1400 B.C. If the Phaistosdisk comes from the period in which Phaistos was the seat of an independent kingdom,
its date, according to Palmer's chronology, would be c. 1700-1600 B. C. I t remains to be
seen, of course, whether the language of the disk is Luvian. And until the disk has been
read, there will always remain the possibility that it is a forgery.
lA.J.A. 58 (1954), p. 88, col. 2.

T H E PROBLEM OF

PUNCTUATION

15

cet usage au profit de la 'scriptio continua' de l'äge classique".
I t is extremely curious that the same change took place in Latin, but after
the classical age.
The practice of word-division was standard in Etruscan and it was probably
from this source that it entered into Latin, where it is found in the very
earliest inscriptions, such as the lapis niger and the fibula Praenestina.e The
word-divider is regularly found on all good inscriptions, in papyri, on wax
tablets, and even in graffiti7 from the earliest Republican times through the
Golden Age and well into the Second Century.
As Revilo P. Oliver has pointed out, "that interpuncta were thought to be
virtually an element of the Latin alphabet may be seen from PSI 743 (I/II),
a curious fragment in which a Greek text has been transliterated in Latin
characters, and the words accordingly separated by interpuncta".8 Seneca
regarded the use of interpuncts as a characteristic of the normal writing of
Latin: "nos etiam cum scribimus, interpungere adsuevimus". 9 The remark
is made in passing to point out the contrast between Greek and Latin temperament, i.e. Greeks use scriptura continua, Romans interpungunt, distinguishing their words clearly even in writing. The reference is at least to the
use of word-dividers; it is possible that Seneca was also thinking of punctuation for sense. 10
Throughout these periods the word-divider was a dot placed half-way between the upper and the lower edge of the line of writing. That the original
form was a vertical line may be inferred from the Phaistos-disk. With alphabets
which used the letter I, to avoid confusion with that letter, the vertical line
was broken to form three dots arranged vertically, as in the oldest Etruscan
inscriptions and on the so-called lapis niger in Latin. The simplification to
two dots and then to a single dot was an obvious development. Two dots
seem normal on the fibula Praenestina (c. 600 B. C.), with three as an exception. The evolution from two to one is beautifully illustrated by the two parts
of the tabulae Iguvinae11 where the two dots divide words in the retrograde
abandonne

• The Duenoa-bowl, with its retrograde characters, is an exception, but the words have
not yet been satisfactorily read, so that we cannot guess whether the absence of the
interpunct could be explained in terms of great antiquity, foreign origin, semi-literacy,
or religious convention.
7
Good examples are C. I. L. IV. 1893 and 1894, which cannot be much earlier than
the destruction of Pompei.
8
Τ. A. P . Α., L X X X I I (1951), pp. 241-242.
»Ep. Mor. IV. 11. 11.
10
The latter meaning is taken for granted by Walter C. Summers in the notes ad loc.
in his Select Letters of Seneca (London 1910).
11
The limits of the dating of the tabulae Iguvinae are 300 to 89 B. C., according to
Tabulae Iguvinae, editae a Iacobo Devoto (Romae, 1940), pp. 51-52. cf. Poultney, Bronze
Tables, pp. 200-289.

16

T H E PROBLEM OF P U N C T U A T I O N

text in the old alphabet, but the single medial point is used in the later text,
written left-to-right, with the new alphabet. The single medial point became
the normal word-divider in Etruscan (e.g. the Liber Zagrabiensis, c. 50 B. C.-50
A. D.) and became standard in Latin at a very early date, subject only to a
few early variations in shape, evidently for decorative purposes.
The Gordons have described the shape of the interpuncts in early inscriptions: "Although our collection includes some Republican examples of crosses,
crosses within squares, and squares and triangles cut in outline, Part I of the
Album has none of these types of punctuation. The regular punctuation
exemplified here consists for the most part of cut-out, down-pointed triangles,
and a smaller number of cut-out ' c o m m a s ' . . . . Both triangles and commas
are generally placed at mid-height, but they show appreciable variation
in shape, direction, position, and comparative size, often within the same
inscription." 12
As a rule, interpuncta are used simply to divide words, except that prepositions are only rarely separated from the word they govern, if this follows
next. Although interpuncta used as word-dividers greatly increase the legibility of the text, they are not punctuation in the sense in which I use the term.
In that sense, the medial point cannot be used as a sign of punctuation until
it is no longer used to divide words.
The regular use of the interpunct as a word-divider continued until sometime
in the Second Century, when it began to fall into disuse, and Latin was written
with increasing frequency, both in papyrus and on stone or bronze, in scriptum
continua. As Professor Oliver states, " I t was evidently during the Second
Century that there took place in Latin Buchwesen one of the most astonishing
cultural regressions of ancient history. Within that century interpuncta and
regular punctuation disappear, apices become rare and sporadic, and lines
become solid blocks of scriptura continua. For this amazing and deplorable
regression one can conjecture no reason other than an inept desire to imitate
even the worst characteristic of Greek books." 13 So far as I know, no study
has been made to determine at what dates the scriptura continua was an in12

Contributions, p. 183.
Τ. A. P. Α., L X X X I I (1951), p. 242. Professor Oliver further states: "Interpuncta
survive in the fragment De Servio Tullio (POxy. 2088) of the Second Century. A good
specimen of the new style is PRyl. 473, a relatively luxurious book which was reused for
accounts in the Third Century, and is therefore assumed to have been produced in the
Second; it has solid lines, no punctuation at all, and only occasional apices. E v e n in a
grammatical treatise, PMich. 429 (II/III; cf. Lowe, Codices, §212), in which one might
suppose clarity t o have been particularly desirable, punctuation disappears except
where it is necessary to distinguish a vowel or diphthong under discussion from the surrounding text. Of course, some sporadic punctuation is to be found in later manuscripts,
but, so far as I know, no attempt was ever made to abandon the scriptura continua or to
return to the fine lucidity of Augustan standards."
13

T H E P R O B L E M OF

PUNCTUATION

17

novation in good writing, became an accepted fashion, and then became normal. The use of the word-divider appears to decline steadily through the Second
Century. It never entirely disappeared, and very late texts can be cited which
use the interpunct regularly,14 but there must have been a time at which
this usage was recognized as intentional archaism.
In this dissertation, my purpose is to deal with the classical period, and
hence with the punctuation that was used in texts in which words were divided
by interpuncts. I shall use later evidence only where there is some reason
to believe that the punctuation is a survival—or deliberate revival—of what
was prevalent before the interpunct went out of general use.
2. Hederae distinguentes,
which appear only on inscriptions, are normally
merely ornamental elaborations of the interpunctum. I shall consider only
the rare instances in which they appear with regular interpuncta and seem
to indicate a stronger division.
3. The apex and l-longa are, strictly speaking, phonetic signs, but the fact
that they are not used for every long vowel15 does make them punctuation
in an extended sense, especially when they appear on inflections and so do
show syntactical relation. They are not, however, punctuation in the restricted sense, although their use may frequently make clear the meaning
that might not otherwise be immediately apparent without a mark of punctuation in the strict sense. I therefore exclude apices as phonetic symbols;
an apex—or, if you prefer, a mark exactly like an apex—is used as a sign
of punctuation, usually placed above the interpunct, and this, of course,
I shall discuss.18
4. A space may be as much a mark of punctuation as a written sign, and
was so used in both inscriptions and literary texts.17 This I shall consider,
but I shall limit myself to divisions within paragraphs. The common practice
of dividing Latin texts into paragraphs according to subject matter, and
indicating the paragraphs by beginning them on a new line, usually extended
14 E . g. P. Oxy. 871, which, according to the editors, may be as late as the Fifth Century,
and thus later than the great vellum codices of Vergil, regularly uses interpuncta. It is
a fragment, perhaps of a treatise or diatribe, that deals with a moral subject.
15 cf. Quint. I. 7. 2: ut longis syllabis omnibus adponere apicem ineptissimum est,
quia plurimae natura ipsa verbi, quod scribitur, patent, sed interim necessarium, cum
eadem littera alium atque alium intellectum, prout correpta vel producta est, facit.
, β I note in passing that the mark of punctuation which exactly resembles an apex
is probably the source of some (not all) apices that we find erroneously placed over the
first or last syllable of a clause or sentence. This, I think, should be taken into consideration in any future study of the use of apices.
17 cf. Robert Seymour Conway's preface to his edition of Livy (Oxford, 1914), §51:
"Vestigium antiquitatis, quod, quoad incorruptum mansit, vix est cur Livio ipsi abnegemus, in Puteano servatum est, cum persaepe intra vola narrationis spatium iv vel plurium
(interdum etiam x) litterarum vacuum reliquit."

18

T H E PROBLEM OF P U N C T U A T I O N

into the left margin, is a different topic. And naturally, I shall not go into
the division of texts into numbered paragraphs (capita,),1* such as Cicero
takes for granted in the text of laws19 and such as are found in the Lex Gallia
Cisalpina of 705/49 (C. I. L. I 2 , 592) and the Lex Coloniae Oenetivae of 710/44
(C I. L. I 2 , 594).
5. I shall not be concerned here with the lectionary signs (signa critica)
that the Romans borrowed from the Alexandrians. I t will suffice to mention
that these may be found in Reifferscheid^ edition 20 of the Reliquiae of Suetonius, and that the most recent listing of them is by Karl Büchner in his contribution to the Atlantis Verlag's Geschichte der Textilberlieferung (Zürich,
1961), Band I, pp. 329f.
In this dissertation I shall use the following procedure in reproducing texts.
All inscriptions published in the C. I. L. are cited from that work; others
from the publication in which they are most readily available (e.g. from the
Annie iSpigraphique rather than the original publication, unless there is a
difference in the text that is significant for our purpose here). For each inscription I have reproduced or summarized an indication of the object on
which the text appears (e.g. "tabula marmorea" or "cippus marmoreus" or
"basis quadrata" or "urna marmorea"), but have omitted all details concerning its ornamentation, discovery, and present location. (Inscriptions lost
before they were examined by modern epigraphers have been generally disregarded, since early transcriptions can not be relied upon for the details of
punctuation.) The volume of the C. I. L. indicates whether a given inscription
is Roman, Italian, or provincial; a more precise statement of the place where
the inscription was set up would have no significance for the present study.
For inscriptions taken from sources other than the Corpus, I have stated the
region in which they were found. Where the editors have dated the inscription
on the basis of either epigraphic or historical evidence, I have quoted or summarized their dating.
Papyri and manuscripts are quoted from the source that contains the most
complete reproduction of the text available to me.
Except where otherwise noted, I have reproduced the whole of the text
under consideration, but I have omitted without comment additions made
on the stone at a later time or inscriptions appearing on other sides of the same

18

cf. Gordon, Contributions, pp. 151 — 155.
' E.g. Cie. De leg. Agr. III. 2. 4.: "Caput est legis quadragesimum, de quo ego consulto",
implies that the hearers had somehow before them copies in which the paragraphs of the
law were clearly numbered for reference.
20
C. Suetoni Tranquilli praeter Caesarum Libros Reliquiae; edidit Augustus Reifferscheid. Lipsiae in aedibus Teubneri, 1860, pp. 137-144.
l

T H E PROBLEM OF P U N C T U A T I O N

19

monument and, in the case of papyri and manuscripts, corrections and marginal annotations by later hands.
All texts are reproduced line for line as they appear in the original, but it
was not feasible to make the right hand margins regular or to show variation
in the size of the letters used for each line. I have reproduced all punctuation,
apices, and I-longae, but have not attempted to reproduce ligatures or to show
letters which are somewhat taller than others for merely decorative reasons
(as Τ and Y frequently, and other letters occasionally, appear in some styles
of lettering).
In the interests of clarity, I have taken no notice of broken letters in inscriptions or partly preserved letters in papyri; where these seemed certain,
I have reproduced them as part of the text without indication, and where
they were mere traces, as part of the restoration. We are here interested
exclusively in punctuation, and any attempt to reproduce such epigraphical
and palaeographical details in printed copy would merely confuse the eye
unnecessarily. Restorations, which, unless otherwise indicated, are those of
the editor of the text, or, where more than one restoration is proposed, the
restoration which the editor appears to prefer, are here given in italicized
lower-case letters. Square brackets are used only in a few cases where it
seemed necessary to show the edge of the preserved stone or papyrus. I n a few
instances I have thought it desirable to add a transcription of the text with
modern capitalization and punctuation; in these transcriptions, restorations
are within square brackets, and corrections are shown by italics.

II

THE TESTIMONY OF THE GRAMMARIANS

I t is difficult to say when the system of punctuation in which we are here
interested was introduced or first came into use. As we have said, punctuation
to separate words appears in the earliest Latin inscriptions and was standard
usage at least from the time of the so-called lapis niger. The practice of dividing
words from one another could at any time have suggested the division of a sentence into groups of related words forming a syntactical unit or otherwise so related that they could be set off from the rest.
The earliest form of punctuation for sense appears to have been a blank
space. Clear instances are found in the well-known Senatus consultum de
Bacchanalibus, which can be precisely dated to 186 B. C. Of other marks
of punctuation, we can only say that Cicero mentions them in passing1 and
that we find them in use in the time of Augustus, evidently as an established
practice and with no suggestion that they are a novelty; there are a few examples in short texts which may come from the last years of the Republic,
while the paucity of long texts, whether on bronze, stone, or papyrus, from
that period prevents us from determining whether or not they were in general
use.
I t is natural to associate the development of systematic punctuation for
sense with the use of apices and other devices to facilitate the reading of Latin
texts, which are believed to have come into use about the time of Sulla, i.e.
around 80 B. C. This general likelihood is a little supported by the consideration
that the most common mark of punctuation, the virgula (/) has the same form,
and is sometimes of the same size, as the apex.
I t is not likely that punctuation came into use without being discussed
' I n De or. III. 44. 173 he mentions rhythmical clausulae (versus in soluta oratione,
explained as numeri quidamJ as a means whereby clausulae are, so to speak, made orally
interpwnctae, the rhythm replacing the librariorum notae. This suggests that punctuation
for sense was common, if not universal, in Latin books in Cicero's time. But it would be
temerity, I think, to argue that the reference proves that punctuation was in common use
in Roman books by September of 91 B. C., the dramatic date of the dialogue.

THE TESTIMONY

OF T H E

GRAMMARIANS

21

by some of the authors who concerned themselves with the proper writing
of Latin. We are certainly entitled to conjecture that Lucilius ( f l 0 3 B. C.)
probably said something about punctuation by blank spaces (which was
certainly known to him), if not by special signs (which could have been known
to, or even invented by, him, even though no example has survived from his
time), in the thirty books of Saturae of which only a thousand lines, or less,
have survived.2 We know that he took an extraordinary interest in devices
to make Latin writing more perspicuous and immediately intelligible, going
so far as to urge that a genitive singular of the second declension (e.g. pueri)
be differentiated by spelling from the nominative plural (puerei) to assist
the eye, although there was no difference in pronunciation.3 We owe this
knowledge to the almost fortuitous preservation of a single fragment, and
it is only reasonable to suppose that such a mind would have taken an interest
in punctuation for sense as another means of assisting the reader. It is not
unlikely, therefore, that he made some reference to it somewhere in the twentynine thousand lines or so that are lost.
It is entirely possible that Varro discussed the subject somewhere, if not
in the De Lingua Latina, where there appears to have been no place for it
according to the necessarily sketchy outline of the contents of the lost books
reconstructed by modern scholars,4 then in some other work, for in his De
Sermone Latino5 he recommended the use of a special mark (nota I transversa)
to call attention to short syllables that occur in verse at points at which the
metre leads the reader to expect a long.
Lucilius and Varro are the only two writers of the Republic whom we can
identify as having discussed subjects closely related to punctuation for sense,
but we could expect the subject to have been treated by some, if not by all,
of the numerous Latin writers who dealt, more or less professionally, with the
orthography and grammar of the Latin language during the late Republic,
when such punctuation seems to have come into use, and during the early
Empire, when we know that punctuation was regularly used in books and
frequently used in long inscriptions. It is true, however, that the comparatively
numerous fragments of the grammatical writers of the two periods, collected
2

There are 120 numbered lines in Lachmann's edition (Berlin, 1876), but m a n y of
these consist of a single word.
3
Lines 317-19 Lachmann. Lucilius m a y have stipulated that I-longa was to be used
in the genitive singular; cf. Velius Longus quoted by Lachmann ad loo. A difference in
pronunciation cannot be presumed from 322-23.
4
For such an outline, see Roland G. Kent's introduction to his edition of Varro in the
Loeb Library, pp. ix-xi.
5
Frag. 66 Goetz and Schoell ( = 46 Funaioli). I t is not clear whether Varro's discussion
of written metrical accents in Greek (Frag. 84 G-S = 282 F) was to have any application
to Latin.

22

THE TESTIMONY

OF T H E

GRAMMARIANS

by Funaioli 6 and Mazzarino, 7 contain no reference whatsoever to the subject.
Since it is scarcely conceivable that none of these writers mentioned the matter,
and is probable t h a t many did, this silence will seem extraordinary until
we remember that most of the fragments of these writers were preserved by the
later grammarians, such as Diomedes (late Fourth Century), Charisius (late
Fourth Century), and Priscian (early Sixth Century), whose works have
come down to us more or less intact and are thus, in a sense, responsible for
the loss of the far more valuable treatises of the grammarians of the classical
age. Our fragments of the earlier writers, therefore, have not been preserved
by chance and at random, but by selection on the part of the later writers.
I t is a reasonable inference, therefore, t h a t those writers preserved by quotation
only passages which they deemed relevant to the 'modern needs' of a decaying
civilization, and therefore disregarded as otiose references to a system of
punctuation that had become obsolete long before their day.
The late grammarians, whose works have been collected and edited by Keil
in an edition which is still accepted as the standard 8 and, it appears, is not
likely to be superseded in our time, 9 do not by any means neglect the question
of punctuation. On the contrary, they lay down, with really astonishing
agreement and unanimity, definite rules of punctuation, which, they imply,
are, or should be, universally used in the writing of connected Latin discourse.
But the rules they give do not pertain to the specifically and distinctively
Latin punctuation in which we are here interested; on the contrary, the rules
enjoin the use in Latin of a system developed for Greek.
Thompson mentions the regular system of Greek punctuation developed
in the schools of Alexandria, invented by Aristophanes of Byzantium (260
B. C.), using "the full point with certain values in certain positions (Θέσεις):
the high point (στιγμή τελεία), equivalent to a full stop; the point on the
line (ύποστιγμή), a shorter pause, equivalent to our semicolon; and the
point in a middle position (στιγμή μέση) an ordinary pause, equivalent
• Grammaticae Romanae Fragmenta, collegit Hyginus Funaioli. Vol. I. Lipsiae in aedibus
B. G. Teubneri, 1907. (This volume is devoted to fragments from the Republican period;
the death of the editor prevented continuation of this edition.)
7
Grammaticae Romanae Fragmenta Aetatis Oaesareae, collegit Antonius Mazzarino.
Vol. I. Augustae Taurinorum in aedibus Loescheri, 1955. (This volume covers the period
from the accession of Augustus to the death of Nero, and includes a few addenda to
Funaioli; the second volume has not yet appeared.)
8
Except for Charisius, for whose work an incomparably superior edition b y C. Barwick
appeared in the Bibliotheca Teubneriana (1925).
9
Keil's Grammatici Latini is now being reprinted, without change and presumably
without addenda or corrigenda, by the Olms Verlag in Hildesheim. With all due respect
to Keil, whose great work was published in 1855-79, with a supplement in 1923, it seems
extremely improbable that new editions could not effect some improvements in the texts.
For one example, see below, p. 24., n. 15.

THE TESTIMONY OF THE GRAMMARIANS

23

to our comma". 10 He stated: "The punctuation of Latin MSS. followed in
some respects the systems of the Greeks. . . . From the Latin grammarians
we know that they adopted the Greek system of punctuation by points (Θέσεις,
positurae), to which they gave the titles of 'distinctio finalis', 'subdistinctio',
and 'distinctio media'; but in practice we find that the scribes11 used the points
without consistently adhering to their meaning."
Of the dozen or more grammarians who discuss the system of punctuation
(positurae or distinctio12·), it will suffice here to quote two of the more extensive
treatments.

10

Edward Maunde Thompson, Introduction to Greek and Latin Palaeography (1912),
p. 69.
11
Thompson's attribution to the scribes of the punctuation in our early parchment
codices is certainly wrong in many cases and doubtful in almost all others. I n these codices, beginning with the fragmentary schedae Fulvianae of Vergil and the codex Palatini, which may be the oldest of the complete or approximately complete codices (cf.
Oliver, Τ. A. P. Α., L X X X I I (1951), pp. 251-54), the writing is in scriptura continua
into which almost all of the punctuation has been fitted. Punctuation in a space left
by the scribe is probably his work, but points fitted into scriptura continua probably are
not and are, in any case, too small to show differences in ink or style of writing. The subscription to the Mediceus of Vergil, "Turcius Rufius Apronianus . . . legi et distincxi
codicem" and, a little later, "distincxi emendans," and other colophons recording t h a t
someone "distinxit et emendavit" (see O. Jahn, Berichte . . . der k. Sachs. Gesellschaft der
Wissenschaften, Phil.-Hist., 1851, pp. 327-72), strongly suggest t h a t in the Fifth Century,
and perhaps earlier, codices were delivered to their purchasers in scriptura continua
with little or no punctuation of any kind, and t h a t the purchaser was expected to supply
the distinctiones just as the purchaser of a French broche today is expected to cut the
leaves. I n any event, the person who distinxit appears to have been principally concerned
with inserting points to show the division of words where the scriptura continua would
be most misleading and a few apices to differentiate words. The Mediceus and Palatinus,
for example, show no signs of systematic punctuation for sense, whether by the scribe
or by a later owner of the book. Codices which do show a system approaching t h a t prescribed by the late grammarians—notably the codex of Juvenal written about 600 A. D.
which I mention below—have punctuation which the editors (in this case, C. H. Roberts)
positively identify as the work, not of the scribe, but of a second hand. For another example, see the papyrus codex (P. Ryl. 477) of Cicero, Div. in Q. Gaec., written in good halfuncials of the Fifth Century; according to E. A. Lowe (Godd. Lat. Antiq., II, 226), the
punctuation is entirely the work of a second hand, although the original scribe may be
responsible for some of the few marks over long vowels (e.g. contra me).
12
cf. Sergii De accentibus (in Donatum) (Keil IV, 482): cum distinctio species sit positurae, tamen abusive pro ipsa positura, hoc est pro ipso genere, accipimus distinctionem.
nam cum sit codex emendatus distinctione, media distinctione, subdistinctione, dicitur
tamen codex esse distinctus.

24

T H E TESTIMONY

Dosithei Ars Orammatica

OF THE

GRAMMARIANS

(Keil V I I , 428):

Distinctio est silentii nota cum sensu terminato. huius autem signum est punctum
supra versum positum. 1 3 subdistinetio est diuturnitas quaedam temporis differens orationem ad sententiae qualitatem. 1 1 huius autem signum est punctum sub versu positum,
u t est illud,
et si f a t a deum, si mens non laeva fuisset,
impulerat ferro Argolicas foedare latebras;
Troiaque nunc staret, Priamique arx alta maneret. 1 5
non enim similiter u t in distinctione silentium interpositum tacere permisit. media vel
mora est silentium legitimae distinctionis subdistinctionisque medium obtinens locum,
quae hoc solum servat officium, u t legentis spiritum levissima respiratione refoveat et
nutriat. sic enim pronuntiando reticere quia debet, quoad spirat, quia spiritus ipse a
defectione vincatur, deinde vires resumat. multae autem causae mediae huius lectionis,
primum ne confundantur quae dicola vel tricola ponuntur; deinde u t actus verborum
emineat et luceat, qui ex aliquo venit affectu vel indignatione seu misearatione conlata,
aut certe quadam artatione sermonis quae emfaticos a poetis * . siquis itaque sine media
spiritus suspensione pronuntiaverit
aut hoc inclusi in ligno occultantur Achivi
aut haec in nostros fabricata est machina muros
inspectura domos venturaque desuper urbi
aut aliquis latet error equi ne credite Teucri,
confunditur ratio compositionis generali nomine ligni machinae equi. et
lectumque iugalem,
quo perii, super imponam:
hoc enim voluit intellegi Dido non esse lectum iugalem, quo perierit. immorandum est
ergo et respirandum post 'iugalem' et sic inferendum cum νποκρίοεως affectu 'quo perii.'
distinguere autem oportet ante similitudines, quas Graeci parabolas vocant, et ante
redditas ανταποδόσεις et siquando a persona da personam transitus erit factus, et ante
'aut' coniunctionem et ante casum vocativum et ante 'sed' et ante 'quoniam' et ante
interrogativa, 1 * u t
13

cf. [Sergii] Explanationum in Donaium liber I (Keil IV, 533-534): ubi enim plenus
est sensus, hoc est ubi f i t clausula dictionis cuiuslibet, punctum ad caput litterae ponimus.
14
cf. Diomedis Artis Grammaticae liber I I (Keil I, 436-439): subdistinetio est silentii
nota legitimi, qua pronuntiationis terminus sensu manente ita suspenditur ut statim id
quod sequitur succedere debeat. . . . subdistinguendum enim est pro voluntate dicentis.
15
I quote verbatim from Keil's edition, on which I am necessarily dependent. Dositheus
(assuming t h a t he is the author of this work; cf. Schanz-Hosius §836) undoubtedly wrote
the quotation with his own punctuation, including, of course, the punctum sub versu
positum t h a t he is here illustrating, and the appropriate points in the later quotations.
W h a t traces, if any, of this punctuation are preserved in the manuscripts cannot be
determined from Keil's edition.
16
As will be obvious from the context, what Dositheus means by interrogativum a
moment later is a vocative apostrophe or interjection; if we knew hie punctuation, we
could tell whether he means the same thing here or is referring to the interrogatives
(in our sense of the word) qui and unde.

T H E TESTIMONY

OF T H E

GRAMMARIANS

25

quis deus, ο Muaae, qui nobis e x t u d i t a r t e m ,
u n d e n o v a ingressus h o m i n u m experientia oepit?
post interrogativa, u t
Musa, mihi causae m e m o r a .

Cassiodorii De Orthographien (Keil VII, 145 — 146):
Illud e t i a m vos m a g n o p e r e credidi commonendos, u t distinctiones s e n s u u m sollicita
m e n t e perquirere ac ponere debeatis, 1 7 sine q u i b u s n e q u e legere q u i c q u a m c o m p e t e n t e r
n e q u e intellegere praevalemus. scire a u t e m d e b e m u s D o n a t u m a r t i g r a p h u m de posituris
i t a tractasse, u t n o n ibi o r d i n e m sed virtutee e a r u m potius exprimere v i d e a t u r . n a m si
distinctionum seriem p e r g r a d u s cognitos sequeretur, p r i m o p l e n a m , deinde m e d i a m
n e q u a q u a m ponere potuisset; nec i n d e magis inciperet, ubi p r a e d i c t a e p o s i t u r a e ad f i n e m
tendere c o m p r o b a n t u r . sequitur subdistinctio, deinde media, q u a e p l e n a s semper praecedunt p o t i u s q u a m s e q u u n t u r . sed sicut visum constat esse doctissimis h u n c debemus
ordinem custodire, u t primo de subdistinctione dicamus, q u a e ibi semper a p p o n i t u r , ubi
in c o m m a t e sermo suspensus a d h u c r e d d e n d u s esse cognoscitur, u t a r m a v i r u m q u e cano,
ubi totius operis s u m m a conclusa est. a r m a enim (ancilla) Vergilius v i r u m q u e d i c t u r u s
est. sed q u o n i a m 'cano' respicit a d u t r u m q u e , hie, id est in ' a r m a , ' subdistinctio recte
p o n e n d a est. m e d i a m vero a d f i g e n d a m c o n s t a t in commate, c u m nullus sermo deest,
sed g r a d a t i m t e n d i t ad p l e n a m , u t
dividimus muros, et moenia p a n d i m u s urbis.
plena est a u t e m , ubi f i n i t u r p e r f e c t a sententia, u t est illud,
t a n t a e molis erat R o m a n a m condere g e n t e m .
subdistinctionem vero vel m e d i a m n o n credas i p s u m ordinem semper t e n e r e q u e m diximus, u t subdistinctio p r a e c e d a t e t m e d i a c o n s e q u a t u r ; sed p r o r a t i o n e s u p r a d i c t a locis
congruis a p p o n u n t u r , u t distinctiones istae bene positae sensum n o b i s lectionis evidenter
aperiant. n a m si aliter distinguas, sine d u b i t a t i o n e c u n c t a confundis. h a s vero distinctiones
seu posituras D o n a t o t e s t a n t e Graeci thesis vocant. periodus est a u t e m p e r l o n g u m plenae
sententiae d u c t a pausatio, cuius p a r t e s s u n t cola et c o m m a t a . 1 8

Such, then, is the system prescribed by the grammarians. So far as I know,
there is no Latin manuscript which embodies this system of punctuation.
There are some approximations to it, of which an interesting example is provided by a page from a parchment codex of Juvenal found at Antinoe in
Egypt and published by C. H. Roberts, who assigns the date of c. 500 A. D. 19
This manuscript was written in uncials and scriptum continua; a second,
but contemporary, hand supplied punctuation, marks of quantity (macron
17
cf. [Sergii] Explanationum
in Donatum liber I (Keil IV, 533-534): o m n i s enim res
initium h a b e t , s e q u e n t i a m e t clausulam a u t , u t dicas hie, i n i t i u m a u g m e n t u m s t a t u m
(declinationis finis sic est positio).
18
cf. Sergii De Accentibue (in D o n a t u m ) (Keil IV, 482): v e r u m m e m i n e r i m u s in p r o s a
cola et c o m m a t a idem esse e t u n u m significare, in versu vero aliud e t diversum.
11
Journal of Egyptian Archaeology, X X I (1935), 199-209 (with plates).

26

T H E TESTIMONY OP THE

GRAMMARIANS

and breve), and word-accents (acute and grave). 20 There are many interlinear
and marginal notes in both Latin and Greek, and it is obvious that the codex
was a carefully prepared edition for the use of readers whose native language
presumably was Greek. With the exception which will be noted below, the
punctuation consists of a medial point, as provided by the grammarians, and
a point on the line, which serves both as subdistinctio and distinctio. There is
no punctum swpra versum positum to serve as full stop. 21 The following examples
will show the use. I give first the text as punctuated in modern editions and
then the text from Antinoe.
VII. 157-164:
nosse v o l u n t omnes, m e r c e d e m solvere nemo.
"Mercedem appellas? Quid enim scio?" Culpa docentis
scilicet arguitur, quod laevae p a r t e mamillae
nil salit Arcadio iuveni, cuius mihi s e x t a
q u a q u e die m i s e r u m dirus c a p u t H a n n i b a l implet,
quidquid id est de quo deliberat, a n p e t a t u r b e m
a Cannis, a n p o s t nimbos e t f u l m i n a cautus
c i r c u m a g a t tnadidas a t e m p e s t a t e cohortes.

In the manuscript this is:
nosseuoluntomnesmercedemsöluerenemö . 3
mercedemappellasquidenimscio · culpadocentis
scilicetarguitur · quodlaeuaepartemamillae
3 nllsdlitarcadicöjuueni · cuiusmihisexta
quaq-dieimiserumdiruscaputhännubalimplet
quidquididestdequodeliberat · a n p e t a t u r b e
acannis · a n p o s t n i m b o s e t f u l m i n a c a u t u s
circumägatmadidasatempestatecohörtes.

VII. 184-190:
Q u a n t i c u m q u e domus, veniet qui fercula docte
componit, veniet qui p u l m e n t a r i a condit.
H o s inter s u m p t u s sestertia Quintiliano,
u t m u l t u m , d u o sufficient: res nulla minoris
c o n s t a b i t p a t r i q u a m filius. " V n d e igitur t o t
Quintilianus h a b e t s a l t u s ? " E x a m p l a n o v o r u m
f a t o r u m transi. Felix e t pulcher e t acer e. q. s.
20
R o b e r t s s t a t e s positively t h a t all these m a r k s are supplied b y t h e second h a n d ;
the scribe, however, appears t o h a v e l e f t a little room for t h e p u n c t u a t i o n in m o s t places,
a n d a t o t h e r points left, it seems, a slight space where no p u n c t u a t i o n was supplied or
needed.
21
Some medial p o i n t s a r e higher t h a n others, b u t it is d o u b t f u l t h a t a distinction
between these a n d t h e o r d i n a r y medial p o i n t was i n t e n d e d ; certainly t h e higher points
are n o t f u l l stops.

THE TESTIMONY OF THE GRAMMARIANS

27

-

qudnticumq domumu0nietquiferculad0cte
conpönit • uenietqiiipulmentariacondit 3
hösintersümptussestertiaquintilidnö
utmultumdiiosufficiunt · resnüllaminöris
constduitpatriquamfilius · undeigiturtot
quintilianushabetsaltus • exemplanouörum
fatörumtransi · felixetpulcheretäcer

The dot on the line clearly stands for a full stop; it is also used, however,
to separate the two parts of an alternative question (163) and to separate
the two parallel clauses that begin with veniet (185). The use of medial points
instead of a full stop in 189 and 190, and also in 158, probably indicates that
the question and reply were regarded as forming part of the same unit of
thought. With these exceptions a point on the line appears where modern
editors place a period, except at the end of 185, where it may have been omitted
by oversight or have been effaced by damage to the parchment.
What makes this particularly interesting is the presence of the two marks
"3" in the passages which I have cited. These are regarded by the editor as
lectionary signs, but they could also be old punctuation of the kind with which
we shall deal below—marks that somehow survived to even this late text.
The one in 157 is superfluous unless it was added to emphasize or strengthen
the full stop that appears before it, but the one in line 185 may replace the
point which, as a sign of the full stop, one would expect to find in that place.
One of the few examples of a superior sign that corresponds to the grammarians' full distinctio, may be found in P. Oxy. 1379, written in mixed uncials
which may be as early as the end of the Third Century. It contains Livy,
I. 5.7-6.1. I give the words that precede and follow the text, so that the
sentences may be complete.
Romulus non cum globo iuvenum—nec enim erat ad vim apertam par—sed aliis alio
itinere iussis certo tempore ad regiamuenirepastoribus
adregemimpetumfacit
etadomonumitorisalia
comparatamanu · adiuuat
remus · itaregemoptrun
cat > numitorinterpri
mumtumultumhostes
inuasiseeurbematque
adortosregiamdictitans
cumpubemalbanamin
arcempraesidioarmisque
obtinendamauocasset
• postquamiuuenesperpetra
tacaedepergereadsegra
tulantisuidit. extemplo

28

T H E TESTIMONY OP THE

GRAMMARIANS

aduocatoconcilio . scele
rainsefratris · originem
nepotumutgeniti
ut educati, ut cogniti essent, caedem deinoeps tyranni seque eius auctorem ostendit.

Noteworthy are the use of points on the line to set off an ablative absolute
construction (though not a similar construction a few lines earlier, perpetrata
caede), of medial points to show that Remus is not the subject of obtruncat
and to show that both scelera and originem are parallel objects of the verb
ostendit. We have already called attention to the full stop after obtruncat.
This is the only end of a complete sentence in our fragment (some modern
editors place a full stop after Remus, but others use a colon).
There are, of course, many other examples of manuscripts from late antiquity
that more or less approximate the rules given by the grammarians. These
suffice to show that the grammarians' treatment of the subject was not purely
theoretical, but they naturally shed no light on the punctuation that was
used in the great age of Latin literature and of Rome, for which the testimony
of the late grammarians is almost valueless.
The testimony of the late grammarians is of some value, to be sure, for
their teaching that punctuation marks indicate pauses in speaking is doubtless
correct and represents much older teaching on the subject. But the system
that they expound obviously can have no connection with the punctuation
used in the classical age, for the points of which the grammarians speak
cannot have come into use until after the separation of words by interpuncts
had been completely abandoned and all texts were presumably written in
scriptum continua. The doctrine of the grammarians, therefore, must date
from a time when scriptura continua, which appears to have been introduced
into Latin in the time of Hadrian or later, had so completely supplanted
the older and clearer form of writing that no one any longer thought of using
interpuncts to separate words.
I t follows, therefore, that since the statements of classical grammarians who
mentioned punctuation are lost, and the late grammarians describe a totally
different system, our only means of obtaining information about punctuation
in classical Latin is to examine the surviving texts and seek to elicit from them
the rules or conventions which governed the use of punctuation.

III
T H E RES GESTAE

AUOÜSTI

The questions that a systematic study of Roman punctuation must answer,
so far as may be possible, may most conveniently be posed by an examination
of a celebrated inscription of considerable length which contains a fairly
large number of signs that were obviously inserted for the purpose of making
the meaning more perspicuous and thereby facilitating the reading of the
inscribed text.
The Res gestae Augusti, also known as the Monumentum Ancyranum from
the site on which an epigraphic copy of the text was first discovered, is certainly one of the most important, if not the most important, Latin inscription
when considered from the standpoint of the historical information which
it contains, and accordingly, since the time of Mommsen, it has often been
called the titulus primarius or "Queen of Inscriptions" in recognition of its
value as an historical document. 1 Its place in a study of Latin punctuation
1 Since the edition of the Monumentum Ancyranum that appears in the Corpus Inscriptionum Latinarum was prepared in 1873 before casts of the original were available
to the editor, it may be disregarded. I t was superseded by Mommsen's special publication, Res Gestae Divi Augusti ex monumentis Ancyrano et Apolloniensi iterum edidit
Th. Mommsen, Berolini apud Weidmannos, 1883, to which was appended a quarto supplement containing a photographic reproduction of the whole of the Ancyran text. For the
subsequently discovered fragments at Antiochia, see David M. Robinson, "The Res
gestae Divi Augusti as Recorded on the Monumentum Antiochenum," American Journal
of Philology, X L V I I (1926), 1-54 plus plates; and Monumentum, Antiochenum, herausgegeben und erläutert von William Mitchell Ramsay und Anton von Premerstein (Klio,
Beiheft X I X ) , Leipzig, Dieterich, 1927. The text of the Res gestae has been repeatedly
edited and commented upon, but for the purposes of the present study only the following
need be used:
Res Gestae Divi Augusti ex monumentis Ancyrano, Antiocheno, Apolloniensi; recensuit
Concepta Barini (Scriptores Graeci et Latini iussu Beniti Mussolini editi), Romae,
1937. (The best critical edition; excellent photographic plates of large portions of the
Monumentum Ancyranum.)
Acta Divi Augusti; pars prior, Romae, ex officina typographica Regiae Academiae
Italicae, 1945. (The edition of the Latin text is by the editor of the volume, S. Ricco-

30

THE 'RES GESTAE AUGTTSTi'

is equally conspicuous. It is the longest and most authoritative Latin inscription that contains internal punctuation for sense (i.e. in addition to the interpuncts that divide words) now extant.
Given the great contemporary importance of this document, which was
written by Augustus himself and served both to exalt him and to provide
propaganda for the principate as a political institution, we may assume that
the official text, inscribed on two pilae aeneae erected in front of the mausoleum
of Augustus shortly after his death in 14 A. D., was, in all respects, as carefully executed and as accurate an inscription as was ever set up in Rome. We
have to depend on the fragmentary remains of copies set up in three relatively
minor provincial towns of Asia Minor, where the normal language was Greek:
Ancyra (modern Ankara), Antiochia Pisidiae (modern Yalvac, not to be
confused with the far larger and more famous Antioch on the Orontes in
Syria), and Apollonia Pisidiae (modern Oluburlu).2
bono; t h e Greek t e x t was edited b y Niccolo F e s t a , a n d o t h e r scholars collaborated in
o t h e r p o r t i o n s of t h e volume. T h e L a t i n t e x t here given is a composite t e x t , a n d so
slightly m o r e convenient for our purposes t h a n t h a t given b y Concepta Barini, whose
t e x t shows only t h e A n c y r a n readings a n d relegates t o t h e a p p a r a t u s t h e readings of
t h e o t h e r inscriptions.)
Res Gestae Divi Augusti; t e x t e etabli e t c o m m e n t e p a r J e a n Gage (Publications de la
F a c u l t e des L e t t r e s de l'Universite de Strasbourg), Paris, " L e s Belles L e t t r e s , " 1950.
(Useful only because t h e t y p o g r a p h y of t e x t enables one t o see a t a glance where t h e
A n c y r a n , Antiochene, a n d Apollonian t e x t s overlap.)
Wilhelm W e b e r , Princeps, B a n d I, S t u t t g a r t , K o h l h a m m e r , 1936. (Contains, p p .
102-240 a n d 106*-265*, a very extensive c o m m e n t a r y o n t h e Res gestae; although
W e b e r is primarily interested in t h e personal a n d political c h a r a c t e r of Augustus, his
is t h e only c o m m e n t a r y t h a t considers p u n c t u a t i o n in discussing t h e readings and
proposed restorations.)
T h e m a n y o t h e r studies of t h e Res gestae (for select bibliographies, see t h e editions b y
B a r i n i a n d Riccobono) contain nothing of significance on t h e m a t t e r s which I shall
discuss.
1 refer t o t h e t e x t of t h e Res gestae b y t h e p a r a g r a p h s , which a r e indicated in t h e inscriptions a n d a r e n u m b e r e d uniformly in all editions. W h e r e a more precise reference
t o t h e A n c y r a n inscription seems desirable, I use t h e usual f o r m which gives t h e n u m b e r
of t h e pagina followed b y t h e n u m b e r of t h e line (e.g. I V . 3 — t h e t h i r d line on t h e f o u r t h
pagina); m o s t editions show these n u m b e r s also.
2
I t is p r o b a b l y merely coincidental t h a t t h e surviving r e m a i n s of a n inscription presum a b l y set u p t h r o u g h o u t t h e R o m a n world a r e f o u n d in w h a t w a s administratively t h e
province of Galatia. Of t h e three, only t h e f i r s t t w o preserve portions of t h e L a t i n t e x t a n d
so serve our purposes here. I t h a s been conjectured t h a t in Apollonia only t h e Greek
t r a n s l a t i o n of t h e Res gestae was inscribed; it is t r u e t h a t only f r a g m e n t s of t h e Greek h a v e
been f o u n d , b u t all these come f r o m three (or n o more t h a n four) of t h e seven paginae
t h a t , according t o m o s t calculations (see Barini, op. cit., p p . viii f., Riccobono, op. cit.,
p p . 13 f.), were needed t o contain t h e Greek t e x t . I t is h a z a r d o u s t o infer f r o m such evidence t h a t t h e L a t i n t e x t was n o t inscribed a t all—all remains of it could h a v e as easily
disappeared as r e m a i n s of t h e t h r e e or four Greek paginae of which we h a v e no fragments—

THE 'RES GESTAE ATJGTTSTl'

31

Although there was a great deal of speculation and argument, so long as
only the inscription at Ancyra, which is by far the most extensive of the three,
was generally known and studied, concerning the possibility t h a t the Greek
translation and other characteristics might have been local additions, it is now
generally agreed that the inscriptions, both Latin and Greek, were copied
from manuscript copies of an official text prepared a t Rome, and that the
comparatively minor variations in matters of detail between the three inscriptions must be attributed to unintentional inaccuracy on the part of either
the scribes at Rome who presumably prepared many copies of the official
text for transmission to provincial governors or the scribes who (whether
at Rome or in the provinces) made copies sent to the various towns and
cities or the scribes who made the copies that were probably made for the use
of the stone-cutters or the designators who presumably sketched in chalk
the letters to be incised on the stone or, finally, the stone-cutters themselves.
I t is possible, and even probable, t h a t each of the persons who intervened
between the official text and the stone made some contribution to the very
minor discrepancies t h a t are to be found in the preserved inscriptions, but it
seems neither possible to determine, nor profitable to inquire, which hand
was responsible for a given difference in orthography, the use of apices, or
punctuation. 3 None of the observed differences between the inscribed texts
affected the meaning. The most important and conspicuous of them are
merely orthographic or insignificant variations of word order. The text of
Ancyra, for example, has inpensarum and incohavi while the one at Antiochia
has impensarum and inchoavi; the former has quae Marcia appellatur and
and we must further allow for the possibility that at Apollonia the Greek and Latin texts
need not have been adjacent to one another, as they were at Ancyra; ef. David M. Robinson's discussion of the placement of the inscription at Antiochia, op. cit., pp. 22—25 and
Plate V I I - B .
3
A very interesting exception m a y be found in VI. 22-23, where the t e x t reads ClSTeri
qui m l H I · QVOQVE . I N MA^wsTRAfV · C 0 N L E G A E · Γuerunt. The stones are so
fractured that only the first apex is entirely certain (hence the differences in reporting
them by Barini, Riccobono, R a m s a y and von Premerstein, and Weber, p. 164), but it
is quite clear that there was no apex above quoque, where one is certainly needed, since
the meaning must be quöque in magistratu (i. q. in singulis magistratibus); cf. Riccobono
ad loc. and F. E. Adcock, J. B. S., X L I I (1952), pp. 10-12. N o w w e are entitled to infer
that the omission of the apex is an inaccuracy that goes back to copies made in Rome,
for it is clear that the m a n who prepared the official Greek translation understood the
word to be quöque and so ignored it in his translation, thus obscuring a very important
point that Augustus was (disingenuously, of course) trying to make. H a d the translator's
Latin t e x t read QV0QVE, he could not have misunderstood, and had he grasped the
meaning, he would surely have added έκάστοτε to his translation, as Riccobono notes
ad XVIII. 8. The quality of the Greek translation as a whole has been much discussed;
see, most recently, Riccobono's commentary on the Greek t e x t and Nicolö Festa's
animadversiones criticae, appended to Riccobono's edition, pp. 65-75.

32

THE 'RES GESTAE

AUGUSTl'

the latter, quae appellator Marcia.1 There are also some differences in the use
of apices, but the two texts generally coincide.
Since it would be vain to speculate concerning what punctuation appeared
in lost portions of either inscription, I assume that the same punctuation
appeared on both and here call attention only to the very few instances in
which the punctuation differs in passages preserved in both inscriptions.
As is to be expected in every good inscription of the period, the interpunct
is used regularly and systematically throughout to divide words from one
another. As is usual, the interpunct is frequently omitted between a monosyllabic preposition and the noun that it governs, thus presumably indicating
that the preposition was treated as a proclitic. One also finds here, as in other
inscriptions, a curious and apparently inconsistent division of compound verbs
(e.g. I N T E R • ESSENT [ I I I . 3], P E R · F f i C I [IV. 14], but R E F f i C I [ I V . 11]
and P E R F I C I [IV. 16]. I t is in this usage that the two texts most frequently
differ; they agree on P R A E T E R · MISSO (IV. 13), but the, Ancyran text
has P R O F L I G A T A (IV. 13) and I N T E R F E C T O (V. 30) where the other
has P R O · F L I G A T A and I N T E R · FECTO.
I feel certain that the interpunct used within compound verbs on the Ancyran text does not differ in any way from the interpunct normally used
between words; it would be extremely interesting, if David M. Robinson
were correct in identifying on the text from Antiochia a special form of the
interpunct in compound verbs,5 but although the interpunct in the word
Ε · D I D I 0 is certainly crowded between the letters and very small (if, indeed,
it is more than a flaw in the stone), in the two clearest and best preserved
occurrences ( P R O · F L I G A T A and I N T E R · FECTO) the interpuncts, as
shown on the available photographs and tracings, do not, to my eye, differ
perceptibly from interpuncts that separate words.
With the exception of proclitics, as noted above, and a very few omissions
that were evidently mere oversights, the interpunct is regularly placed after
every word, including words that occur at the end of a line or at the end of
a paragraph. The interpunct accordingly precedes signs of punctuation for
sense wherever these occur.
Seven distinguishable signs of punctuation appear in the preserved copies
of the Res gestae, viz. (1) a blank space; (2) a diagonal bar (/); (3) a small
diagonal (similar to an apex) not much larger than an interpunct (') and placed
close to it or even above it; (4) a small diple pointed to the right ( > or occasionally 7 with the top stroke nearly horizontal); (5) a sign that vaguely
4 For a list of the differences between the Latin texts of Ancyra and Antiochia, see
Riccobono, pp. 17-18.
s Robinson, pp. 31, 33.
* Discussed b y Robinson, p. '.13; the placing of this tiny fragment of stone is in any
case uncertain.

THE "RES GESTAE AUGUSTI'

3c

resembles two such diplae placed one above the other with the lower one
curved (3) — and we may regard the sign ζ , of which a single instance is found
in the Antiochene inscription as either a variant of this or as an oddly ornamented > ; (6) a vertical bar with hooks top and bottom (J*) found thrice
on the Ancyran inscription and with minor variations that will be noted
later; and (7) a sign which is at Ancyra a large inclined stroke curved at the
to
P ( ^ ) a n d at Antiochia consistently appears in a distinctly different
form
the only instance of a definite difference in "style" between
the two monuments. 7
Editors, following the lead of Mommsen, do not distinguish among these
various signs. Some, indeed, including, oddly enough, Sandys in his Latin
Epigraphy (pp. 258-276), simply ignore the punctuation of the original;
the others adopt the practice of Mommsen, who used [§] to mark a blank space
and § to replace indiscriminately any of the six written signs. (The only departure from this procedure that I have noted was made by Kiccobono, who
uses ~ to indicate a blank space, but retains § as a substitute for any mark
of punctuation. What is more remarkable, Mommsen in his diplomatic edition
of the text, although he naturally used a blank space to show a blank space
in the original, used a diagonal bar (/) to replace all of the six marks that X
have distinguished. Mommsen seems to have taken it for granted that these
six marks were all equivalent, and in any case he was not much interested
in them: "omnino re non differunt", he says, 8 "pendentque ab arbitrio
scribentis notae hae omnes, Eas si qui erit qui curat, tabulas [sc. photographicas] inspiciat omnia ilia plene repraesentantes."
Since I did not have access to the complete casts of the Ancyran inscription
that are in Berlin (if not destroyed in the European catastrophe of 1945),
Rome, and Ithaca, New York, I have had to rely on the complete photographs
appended to Mommsen's edition, the photographs of a large part included
in Barini's edition, and the casts of a small part of the inscription at the Uni7

The form
appears consistently in the Antiochene fragments, as will be seen by
inspecting the photographs and tracings. If the epigraphers who believe that they can
distinguish in the Antiochene fragments the work of at least four different stone-cutters
(see Ramsay and von Premerstein, pp. 19 f.) are correct, it may be significant that all
four of these adopted a form of the paragraph-sign that differs very distinctly from the
form regularly used in Ancyra. The reader should consult the plates in Ramsay and von
Premerstein's edition without permitting himself to be prejudiced by the editors' attempt
to suggest the sign on p. 22 by using such type as the printer had in his cases—with the
odd result that it there does resemble the sign found in the Ancyran text much more
than what is shown on the plates. This may mean that Ramsay and von Premerstein
noted the difference between the Antiochene and Ancyran forms and decided that the
latter was more "correct," although they say nothing about it and do not mention the
difference which I have noted.
8
p. 190.

34

THE

"RES G E S T A E

AXJGUSTl'

versity of Illinois, supplemented by such hints as occasionally appear in the
apparatus of Barini and Riccobono and in the commentary of Weber. As will
be seen, I have had to remain in doubt at some points where inspection of
the stone or a good cast might have removed the uncertainty. For the text
at Antiochia, Robinson's photographs and especially the tracings given in
Ramsay and von Premerstein seemed to be everywhere sufficient, except
for a few fragments of the stone which are said to be now lost.
The marks of punctuation that I have listed above may be clearly seen
on the stones of the inscriptions, and their shapes are so distinct and different
that it could scarcely be argued that one is a scribal variant or whimsical
deformation of another. I should remark, however, that all marks of punctuation
are more lightly incised than the rest of the inscription. The lines are not so
wide as either the vertical or the horizontal bars of the letters, and they are
not cut so deeply into the stone as either the letters or the normal interpuncts
between words. The marks of punctuation are all thin and shallow marks,
resembling in this respect the apices which are placed above long vowels.
Marks of punctuation, therefore could be more easily effaced than letters.
Slight damage to the surface of the stone could destroy them or enough of
each mark to make it uncertain what its full form was. (As will be noted later,
I have encountered a number of places where the stone, as shown by the
photographs, preserves what are clearly vestiges of some sign but not enough
to enable me to say definitely what is was.) It is even possible that some parts
of these signs were so lightly scratched on the surface by a lapidary seeking
to make a very light and thin mark as to be obliterated by mere weathering.
The blank spaces, instead of being a distinct sign, could be merely spaces
in which the stone-cutter either forgot to incise a mark of punctuation or
scratched it on so lightly that it has now disappeared.9
Although it is possible to regard the small diagonal as merely a variant
9 Professor Oliver believes that he will be able to show in a forthcoming article that
all professionally produced inscriptions were the work of at least three specialized artisans,
a designator (who sketched the letters in chalk on the stone), a lapicida (who cut them
with his chisel), and a miniator (who filled in the varnish, usually red, that made the
inscription easily legible from a distance); and that furthermore the miniator frequently
corrected with his brush, and without recalling the lapicida with his chisel, small omissions
or other errors made by the cutter, e.g. an F cut on the stone could be converted to an
Ε by the miniator simply b y painting the bottom stroke on the stone, and likewise an
incised Ε could be corrected to F b y simply failing to fill in the horizontal stroke cut in
error. Portions of letters painted in rather than cut would probably not be detected by a
purchaser, if the total text was of any considerable length and the letters not very large.
If this was a common procedure in making inscriptions, the probability of the possible
explanation of blank spaces that I have given is considerably increased: the stone-cutter
may have made no mark and the miniator may have painted in a slender mark of punctuation which, of course, was eventually washed off the stone.

THE

'RES GESTAE

ATJGUSTl'

35

of the larger one, used when it was necessary to crowd letters together or added
as an afterthought when the following letter had been cut and there was no
longer space available for the larger sign, and although it is conceivable that
the > is merely the upper part of a 3 of which the rest has vanished, it is
apparent that we are dealing with at least three basic forms, / , 3, and ,
^
(or
, if we prefer the Antiochene form of the sign), no one of which could
be a mere illusion produced by damage to the stone or the result of a mere
oversight by the lapidary.
To determine what distinct values, if any, are represented by the various
signs that I have listed, we must consider their function in the inscribed text.
I shall accordingly proceed to examine the use of punctuation to set off the
various units of the text, viz. paragraphs, sentences, and parts of sentences.
PARAGRAPHS

The inscribed text of the Res gestae consisted of a title, unmistakably set off
by being inscribed in much larger and more formal letters, the text written
by Augustus, which is divided into thirty-five capita, or paragraphs, and a
summary added at the end by another person and forming four additional
paragraphs.
The paragraphs are all clearly set off because the first line of each is, in conformity with the common Roman practice, extended into the left margin
and begins with a conspicuous capital letter (i.e. littera capitalis, a large letter
which marks the beginning of a caput or capitulum). The last line of almost
every paragraph, furthermore, ends far short of the right-hand margin, thus
further separating the paragraph from the one which follows, since the considerable amount of blank space thus left is quite conspicuous.
Although, obviously, the paragraphs are clearly and unmistakably set off
by the edentation and large initial letter at the beginning and the area of blank
space at the end, so that any further indication of paragraph-divisions is pure
supererogation, many of the paragraphs are terminated by the s i g a . ^ , ^
With one possible exception, which I shall discuss below, this sign occurs only
at the ends of paragraphs, and must therefore be regarded as a special sign
which distinctly and specifically means "end of paragraph". It is normally
placed at a considerable distance from the interpunct which follows the last
word in the paragraph. The sign is sometimes placed flush with the right-hand
margin and sometimes about a third or a half of the way between the last
word and the right-hand margin. This variation in its position, incidentally,
makes it impossible to be sure that the sign did not appear after a given
paragraph, if the stone is mutilated in any place where the sign could have
appeared.

36

THE 'BES GESTAE AUGUSTl'

Of the thirty-nine paragraphs (35 plus 4 of the "appendix" of summaries),
the last line of nineteen is either lost in both the Ancyran and Antiochene
inscriptions or so mutilated that it is impossible to determine whether or not
the sign ^ ^ was placed at the end of the paragraph. To these I must add
the last paragraph of all, after which Mommsen and the other editors place a
§, thus indicating t h a t it is followed by some kind of sign, but I cannot discern on the photographs any vestiges of it and so cannot say what shape it
had or where it was placed.
On the remaining nineteen paragraphs, the final line of fifteen (8, 10, 12,
14, 15, 17, 19, 21, 23, 24, 25, 26, 32, 34, Ap. 2) is sufficiently preserved in one
or both texts to show the sign
which concludes these paragraphs and
which is always separated from the last word by a considerable blank space,
appearing sometimes on the right margin and sometimes closer to the last
word than to the right margin. Of these fifteen paragraphs, both inscriptions
preserve the terminal mark of two (21, 26).10 There are also in this group two
discrepancies between the two texts to be noted. The terminal mark of paragraphs 15 and 25 is preserved in the Antiochene text, but does not appear
in the Ancyran. The reason for this is apparent in 15, where the Ancyran
lapidary was just able to complete the words of the last line, with some slight
crowding, in the space available, and so had no room for the terminal sign.
In paragraph 25, the interpunct following the last word is closely followed
by the sign > ; the condition of the stone is such that a
could have appeared later in the line.
Two paragraphs of which the ends are preserved only at Ancyra end with
other marks of punctuation. Paragraph 3, like the 25 just mentioned, has a
> placed immediately after the last interpunct. In paragraph 27, the interpunct
is followed by a considerable blank space and by the large sign ^ . In the first
instance an additional
5 could have been lost, if placed at the right margin. In the second, it seems certain that nothing followed the unusual mark.
In the remaining two paragraphs, both in Ancyra, there is no terminal
mark. Paragraph 11, like the 15 mentioned above, has a final line that the
lapidary was just able to complete at the right margin with no room left for
a terminal sign. At the end of 16, however, there is plenty of room and the
surface of the stone is fairly well preserved, so that it appears that the terminal
sign was simply omitted by an oversight.
I t will thus be seen that the terminal sign
was apparently regularly
employed at the end of paragraphs and intentionally omitted only where there
was no room for it. No reason can be found for the two instances of > and
the occurrence of ^ , apparently in place of

10

1 have already called attention to the variant form of the sign used at Antiochia

ΤΗ 10 ' r e s g e s t a e

augusti'

37

One further occurrence of ^—— remains to be noted; in the Ancyran text
it is used, preceded and followed by a considerable blank space, to divide
paragraph 8 into two parts. (The Antiochene fragment has lost the portion of
the stone on which this sign, if present, would have appeared.) The use
is not an error; it is a very strong mark of punctuation to show a complete
change of subject. In the part of the paragraph that precedes the sign, Augustus
reports the number of Roman citizens enumerated a t each census. After the
sign, he says: "Legibus novis inlatis multa exempla maiorum exolescentia
iam ex nostro saeculo reduxi et ipse multarum rerum exempla imitanda posteris tradidi." This is obviously an entirely different subject, unconnected with
what precedes. I t could properly have formed a separate paragraph, and indeed, but for the clear evidence of the Antiochene text that it did not, we
should infer that the Ancyran lapidary erred when he did not extend Legibus
into the left margin. I t seems proper to conclude, therefore, that the paragraph
sign could be used to show the kind of change of subject that would normally
call for a new paragraph, even when, for some reason (perhaps because the
passage would have made a paragraph of only three lines) the division was not
marked by edentation and a capital letter.

SENTENCES

The style of Augustus, which has been much discussed by the commentators, 11
is simple and direct. He used comparatively short sentences, and the only
room for serious doubt concerning what constitutes a sentence in our text
is found where two parallel clauses are connected by et. There is, however,
some little difference between modern editors in this respect, and I shall
follow Riccobono, who recognizes a few more sentences than does Barini.
In the text as punctuated by Riccobono there are 120 sentences, viz. one
in the title, 115 in the thirty-five paragraphs by Augustus, and four in the
four appended paragraphs. Now since the title is set off by its style of lettering,
thirty-nine of these sentences end with paragraphs, and one, which I have
just mentioned, ends with a paragraph sign, there are in the total text seventynine sentences which should have a full stop as a mark of punctuation to separate them from the sentences which follow.
Of the seventy-nine places in the text at which we would expect to find
punctuation showing the end of a sentence, twenty-two occur at points at
which both inscriptions are defective, i.e. the stone is either missing or so
mutilated that it is impossible to determine whether or not punctuation was
present. Four others (in paragraphs 23, 26 bis, 30) occur at points where the
"Especially Weber, op. cit., pp. 102-240.

38

THE 'EES GESTAE ATJGUSTl'

editors can evidently identify the traces of letters, b u t the photographs available
to me do not show these clearly. The ends of forty-four sentences are clearly
marked by punctuation of one kind or another. And at nine points where
Riccobono (who does not differ greatly from other modern editors in this)
places a full stop, the text shows no punctuation, thus suggesting that the
writer's conception of what constitutes a sentence may not always have corresponded to Riccobono's. To this question I shall return presently.
Of the forty-four instances of punctuation corresponding to a full stop,
seven appear to be merely blank spaces, although in four of these seven instances the stone is so roughened t h a t I am not certain that some lightly incised mark of punctuation could not have been present at one time. The three
instances where the blank space seems certain occur in paragraph 8 after
tria millia, separating t h a t sentence from the following one, which begins
"Tum iterum consulari cum imperio . . ."; in paragraph 27 after privignus erat,
separating this from the following sentence, " E t eandem gentem. . . " and
in the same paragraph after Artavasdi, separating the sentence there ending
from the following "Quo interfecto T i g r a n e < m > . . .in id regnum misi."
Of the thirty-seven places where marks were certainly present (and further
set off by a generous allowance of space before or on both sides of them), in
seven the traces on the stone (at least so far as they can be seen in photographs)
are not sufficiently complete to enable me to determine with any certainty
what the mark was.
At sixteen places the end of the sentence is marked by a diagonal bar, and
for one of these we have fortunately both inscriptions available and in accord:
paragraph 20,12

RIVOS · AQVARVM · COMPLtfRIBVS · LOCIS · V E T V S T l T E ·
LABENTES · R E F i l C I · E T · AQVAM · QVAE · MÄRCIA ·
APPELLÄTVR · DVPLICAVI · FONTE · NOVO · I N R I W M · EIVS ·
INMISSO • I FORVM · iVLIVM · ET · BASILICAM · QVAE · FVIT ·
I N T E R · AEDEM · CASTORIS · E T · AEDEM · SATVRNI · . . .
P E R F f i C I · e. q. s. 13

12

I n these quotations, since we are interested only in t h e punctuation, I do n o t attempt
to indicate w h a t letters are mutilated or missing in one or t h e other inscription, and in
the present quotation I e v e n disregard the fact t h a t t h e order of the words Marcia
appellatur differs. I do n o t show the division into lines since this differs in t h e t w o inscriptions.
13
The marks of suspension in m y quotations, I need scarcely remark, show portions
of t h e t e x t t h a t I h a v e omitted to concentrate t h e reader's attention on t h e essential
grammatical structure in question.

THE

'RES GESTAE

39

AUGUSTl'

In eight instances, the full stop is indicated by the symbol > , and here
again we have one instance in which the symbol is preserved in both inscriptions, 11 paragraph 5:
... .quAM · ITA · ADmmlSTram ut INira · D l f i s · PAVCOS · METV ·
E T · PERICLO · PrAESENTI · CIVITÄTEM · VNIVersam liberarem impensa
et • CVRA · MEÄ · > CONSVLaiwm QVOQVe · TVM · ANNVVM · Et perpetuum mihi DELAiwra non recepi.
There are two instances in which the two texts differ. In paragraph 15, the
Ancyran has:
. . . . HOMINVM · CIRCITER · CENTVM · ET · V I G I N T I · MILLIA · /
CONSVL · TERTIVM · DECiMVM · S E X A G E N 0 S · D E N Ä R I 0 S ·
P L E B E I · QVAE · TVM · F R ^ M E N T V M · PVBLICVM · ACCIPIEBAT ·
D E D I • e. q. s.
The Antiochene text divides the sentences with a > . I n paragraph 8, the
Ancyran has:
. . . IN · CONSVLÄTV" · SEXTO · SfiNSVM · POPVLI • CONLEGÄ · Μ ·
AGRIPPÄ · fiGI · I LVSTRUM • POST • ANNVM · ALTERVM · E T ·
QVADRAGENSIMVM · FfiCi e. q. s.
The Antiochene text uses the curious symbol £ , which appears only a t this
point, and may be only a reversal of the symbol 3, which appears as a full
stop on the Ancyran stone in paragraph 29:
SIGNA · M I L I T A R I A · COMPLVRa per A L I O S · D m C E S ·
DEVICTls hostibuS • REcipeRAVl · E X · HISPANIA · E T
et a cWraATEIS · 3 PARTHOS · TRIVM · E X E R C I T ^ M
SPOLIA · ET · SIGNA · REddere MIHI · SVPPLICESQVE
P O P V L I · ROMANI • P E T E R E · COEGI ·

ÄMIssa
• gallia
· ROMANORVM ·
· AMICITIAM ·

The sign obviously marks a full stop. There is nothing, however, which would
indicate that the division between the two sentences was either stronger or
weaker than the division marked by / or > .
The Ancyran inscription, furthermore, contains a marking which may, perhaps, be regarded as a vertical line with small hooks at top and bottom to distinguish it from an I. There are three instances of this mark, and it will be
noted that each differs quite perceptibly from the others; it is also curious
t h a t they occur in three successive paragraphs.
11

Stin Chapters VI and VII for other inscriptions in which this eign is used.

40

THE

'RES GESTAE

AUGUSTl'

25: I V R A V I T · I N • ME A · V E R B A · T O T A · I T A L I A . S P O N T E · S V Ä
E T · M E · B E Hi QVO · V I C I · A D · A C T I V M · D V C E M · D E P O S C I T ·
I V R A V E R V N T · I N · E A D E M · VERöa pnwiNCIAE · G A L L I A E
H I S P A N I A E e.q.s.
26: . . . E T · E I V S D E M · T R A C T ^ S · A L I I · G E R M Ä N O R V M · POPV/I ·
P E R · LEGÄTÖS · A M I C I T I A M · M E A M · ET · P O P V L I · R O M Ä N I ·
P E T I E R V N T • I MEO IVSSV • E T · A V S P I C I O · DVCTI · S V N T ·
duo • E X E R C I T V S e. q. s.
27: A E G Y P T V M · I M P E R I O · P O P V L I r o M A N I · A D I E C I X
A R M E N I A M · M A I O R E M • I N T E R F E C T O · REGE · EIVS ·
A R T A X E p. η. s.

/

The three signs obviously serve as full stops. There is no perceptible difference
in value between the three slightly different forms of this sign, or between this
sign and / or > , where these are used to make the ends of sentences.
W e now come to the nine sentences in the text as punctuated by Riccobono
which are not terminated by an indication of a full stop on the stone. In one
of these ( I I . 9 of the Ancyran text), inspection of the stone will show that the
failure to put a mark of punctuation before the third occurrence of quo lüstro
in paragraph 8 is explained by the fact that the mark of punctuation would
have had to come at the end of a line which is so crowded that there was no
room for it.
Of the remaining instances, four can be eliminated by reducing Riccobono's
two sentences to a single sentence composed of two parallel clauses joined
by et. Here is Riccobono's text of the four:
8. Senatum ter legi. Et in consulätü sexto censum populi conlegä Μ. Agrippä
egi.
14. Filios meos.. .senatus populusque Romanus annum quin tum et decimum
agentls consules designävit, ut eum magisträtum inirent post quinquennium. E t ex eo die, quo deducti sunt in forum, ut Interessent consiliis
publicis decrevit senatus.
16. I d primus et solus omnium, qui dedüxerunt colonias militum in Italia
aut in provincis, ad memoriam aetatis meae feci. E t postea Ti. Nerone
et Cn. Pisone consulibus.. . militibus.. . praemia numerato persolvi.
17. Quater pecuniä meä iuvi aerärium, ita ut sestertium milliens et quingenties
ad eos qui praerant aerärio detulerim. E t Μ. Lepido et L. Arruntio cos.
in aerarium militare. . . HS milliens et septingentiens ex patrimonio meo
detuli.
I t will be obvious that in each of these cases the two sentences distinguished
by Riccobono deal with the same subject and could have been treated as a

THE 'RES GESTAE

AUGUSTl'

41

single compound sentence having two principal members joined by et. Augustus
was noted for the impatience with which he regarded the subtleties of grammarians and for writing very much as he spoke. The loose junction of such
clauses may be a stylistic weakness, but I am inclined to believe that the texts
justify the tentative conclusion—in the absence of evidence to the contrary —
that the copy of Augustus's work treated the four passages above as composed
of single sentences rather than as composed of two.
There remain, however, four instances in which no such explanation can be
found to account for the omission of a full stop, viz.:
15.

. . .consul undecimum duodecim frümentätiönes frümento privatim
coempto emensus sum, et tribuniciä potestäte duodecimum quadringenös
nummös tertium viritim dedi. Quae mea congiaria pervenerunt ad hominum millia numquam minus quinquäginta et ducenta.

16. Pecuniam pro agris.. . solvi münicipiis. Ea summa sestertium circiter
sexiens milliens fuit, quam pro Italicis praedis numeravi...
25. Mare pacävi ä praedonibus. Eö bello servörum, qui fugerant ä dominis
suis et arma contra rem publicam ceperant, triginta fere millia capta
dominis ad supplicium sumendum tradidi.
26. Omnium provinciarum populi Romani, quibus finitimae fuerunt gentes
quae ηδη pararent imperio nostro, fines auxi. Gallias et Hispaniäs prövinciäs, item Germaniam qua includit Öceanus a Gädibus ad ostium
Albis flüminis pacavi.
I n the first of these passages, the quae congiaria, which refers to the whole
series of previously enumerated "benevolences", could in modern writing
be tacked onto the sentence by some device such as a dash, but it is clearly
disconcerting to find no indication of a break between it and the preceding
words. I n the other examples, we are clearly dealing with two separate, though
not unrelated, thoughts, each syntactically complete in itself, and we are left
to choose between omission by the lapidary and negligence by the writer as
an explanation. The former is a little the more probable, since the inscriptions
are not letter-perfect: the Ancyran stone has aede for aedem, and the Antiochene
shows memoria for memoriam. I f a necessary letter could be omitted by the
designator or ignored by the mason, a mark of punctuation could be omitted
in the same way.

42

THE 'RES GESTAE AUGüSTl'
I N T E R N A L PUNCTUATION

Disregarding punctuation noted by editors but of which I can discern no clear
traces in the photographs, and, of course, disregarding editors' restorations,15
there are in the inscription forty-six clear examples of punctuation within
sentences.16
The only mark that is limited to internal punctuation is a very short diagonal
bar (') which is not much larger than an interpunct and occupies no more
space, being placed immediately beside it or, in two instances, over it. The other
marks of internal punctuation are all found also at the ends of sentences:
I > 3 (one instance). I note however a definite tendency to allow less space
for the I as an internal mark and to place the > above the mid-point in the
line. In addition, a blank space on the stone is also used.
One use of internal punctuation is to separate items in an enumeration.
The two longest series in the inscription will incidentally suggest that the
various symbols were regarded as equivalents:
19. CURIAM · ET · CONTINENS · EI · CHALCIDICVM · TEMPLVMQVE · APOLLINIS · IN PALATIO · CVM · PORTICIBVS · AEDEM ·
DIVI · IVLI · LVPERCAL · PORTICVM · AD · CIRCVM FLÄMINIVM · QVAM · SVM · APPELLÄRI · PASSVS · EX ·
Ν0ΜΙΝΕ · EIVS · QVI · PRI0REM · E0DEM · IN · SOLO · FECERAT ·
OCTAVIAM · /PVLVINAR · AD · CIRCVM · MAXIMVM · AEDÖS ·
IN · CAPITOLIO · IOVIS · FERETRI · 'ETTONANTIS · AEDEM ·
QVIRINI · /AEDfiS · MINERVAE · > ET · I ONONIS • REGINAE ·
> ET · IOVIS · LIBERTATIS · IN · AVENTINO · / AEDEM·
LARVM · IN · SVMMÄ · SACRÄVIÄ · / AEDEM · DEVM · PENÄTIVM ·
IN · VELIA · I AEDEM · IWENTÄTIS · / AEDEM · MÄTRIS ·
MAGNAE · IN · PALÄTIO · FECI ·
15
An interesting case m a y be found in 1.39, where Weber (p. 162*, note 598) includes
a mark of punctuation in his restoration, partly because he thinks indication of a mora
is needed to facilitate understanding of the sentence, and partly because a mark of punctuation appears a t the corresponding point in the Greek version, which has ΧΕΙΡΟΤΟΝΗΘΩ ' Α Ρ Χ Ε Ν Ο Υ Δ Ε Μ Ι Α Ν e. q. s. This, of course, is no proof t h a t a mark appeared in
the Latin text, and, if it were, it would still be impossible to determine what kind of mark
it was. I need scarcely observe t h a t in dealing with t h e Ancyran and Antiochene inscriptions, the presence of punctuation in lost portions could not be determined by measurements of space. When editors still disagree as to whether or not there is room for the
restored mihi in 1.36 (see Barini's apparatus, Riccobono's note, and Weber, p. 169)
measurements of space are not sufficiently accurate (since the letters vary in width) to
indicate the presence or absence of a mark of punctuation.
16
Of course, I continue to follow Riccobono. The amount of internal punctuation
could be increased, a t the expense of the external, by recognizing fewer sentences than
does Riccobono; for example, in paragraph 8, Riccobono treats the three statements
beginning with quo lustro as separate sentences, while Barini treats them as relative clauses
annexed to the preceding sentence. I t seems pointless to argue about such matters here.

T H E ' R E S GESTAE ATJGTJSTl'

43

Αρ. 2. OPERA · FECIT · NOVA ·' AEDEM · MARTIS iovis tonantis et feretri
apollinis D I V I I ^ L I · 'QVIRINI ·' MINERVAE iunonis reginae iovis
libertatis LARVM · 'DEVM PENÄTIVM · TWentatis
matris magnae
lupercal pulvinaR · AD · CIRCVM · CVRIAM · CVM · CHalcidico forum
augustum basilicaM. · IVLIAM · THEATRVM · MARCELLI · PORficwm
octaviam nemus trans i l B E R l M · CAESARVM ·
One is at first sight tempted to distinguish between the three functions of the
punctuation: to introduce a series (·' in Ap. 2), to separate the different items
of the series when they are separated at all ( / and blank space in 19, blank
space in Ap. 2), and to separate the names in the genitive in a series depending
on aedes ( ' and > in 19,' and blank space in Ap. 2). But given the inconsistency
found even here and the great variations found elsewhere, the partial specialization of the signs in this enumeration is probably fortuitous.
A somewhat similar enumeration, this time of dates, is found in paragraph
16:

POSTEA · TI · NERONE · ET · CN • PISONE · CONSVLIBVS · ITEMQVE ·
C · ANTISTIO · ET · D · LAELIO · COS · ET · C · CALVISIO · ET · L ·
PASIENO · CONSVLIBUS · ET · L · LEwiVLO · ET · Μ · MESSALLA ·
CONSVLIBVS · 3 ET · L · CÄNINIO · ET • Q · FABRICIO · COe
MILITiBVS... PRAEMiA.. .PERSOLVI.
The space after Cäninio is clearly an error; it should have been put either
after one of the earlier groups or after the final COS. The symbol 3 clearly
means nothing more than a blank space. (There is no grouping of years indicated,
for the years after item are 748, 750, 751, and 752 Α. V. C.). Consistency would
have required the same punctuation after each pair of consuls.
The other uses of internal punctuation may be classified thus: to set off
ablative absolute phrases, to set off certain relative clauses, to distinguish
between parallel clauses in a compound sentence, to show that two or more
statements are governed by the same verb or that several genitives depend
on one noun although they are separated by intervening phrases or clauses,
to mark a collective apposition, and to perform some miscellaneous functions
which we cannot state positively since only one example of each use is available
in our text.
There are four examples of the ablative absolute, including two statements
of date, viz.:
8. tertiuM • CONSVLÄRI · CVM · IMPERIO · L^STRVM · CONLEGÄ ·
TIB · CAE.sare filio Μeo feci • /SEX · POMPEIO · ET · SEX · APVLEIO ·
COS.
(We should note that punctuation does not appear in the precisely parallel
construction earlier in this paragraph: LVSTRVM · <s0LVS · F E C I · C •
CENSORINo et c ASINIO · COS ·)

THE 'RES GESTAE AUGUSTl'

44

12. ΤΙ · NERONE · Ρ · QVIntilio cOnSulibus • ÄRAM · .. .senatus. . .
CONSAcRANDAM · censuit
Other ablative absolute clauses:
10. quOT) • SACERDOTIVM · ALIQVOD · POST · ANN0S · EO · MORTVO ·
Qui civilis Μotus oCCASIONE · OCCVPAVERAT · > CVNCTA · E X ·
ITALIA · ad comitia mea CONFLVENie mwLTITVDINE · . . .RECEPi
Here, perhaps, the intention was to divide two successive ablative absolutes
or perhaps to mark the end of the relative clause depending on the first.
Our last example depends on an emendation of the punctuation, but this is
fairly certain. The text reads:
4. Laurum de /ASCiBVS · DEPOSVI · /IN · CKPltolio votis quae QV0QVE ·
BELLO · NVNCVPAVERAM soZVTIS ·
Since Augustus did deposit the laurel in the Capitol and since he simply could
not have discharged all his vows in that place,17 we are surely justified in
assuming that the punctuation was mistakenly put before, instead of after,
the phrase in Gapitolio.
Relative clauses:
1. . . . EXERCITVM · PRIVÄTO · CONSILIO · ET · PRIVATÄ ·
IMPENSÄ · I COMPARÄVI · > P E R · QVEM · REM · PVBLICAM...
INLIBERTÄTEM · VINDICÄm
(Traces of the punctuation after comparavi are clearly seen in the photographs,
although Mommsen in his edition marks a blank space at the point.
12.

. . . P A R S praetorum
e T · T R I B V N O R V M plebi..
.O&VIAM · M I H I ·
M I S S A · Ε st in campanlAM
• Q V I · H O N O S · ad hoc • T E M P V S ·
N E M I N I · P R A E T E R TOE · ES< decretus

(The space, which obviously serves to set off the relative clause, is clearly seen
in the Antiochene text.)
16. . . . PRAEMiA · NVMERATO · PERSOLVI · QVAM · IN · REM ·
SESTERTIVM • quater mILLIENS • QWlcitell • IMPENDI ·
21. DON«.. .CONSECRlVI · > QVAE · MIHI · CONSTITERVNT ·
HS CIRCITER • MILLIENS
17

See especially Weber, op. r,it., p. 13G*, note 580.

T H E ' R E S GESTAE A U G U S T l '

45

26. CLAssis mEA. . . A D · Fines cimbroruM • N A V I G A V I T · / Q V 0 ·

NEQVE · TERRA · NEQVE · MARI · QVISQVAM · ROMANVS ·
ANTE · ID · TEMPVS · ADIT · / CIMBRIQVE · ET · CHARYDES...
AMICITIAM · ME AM. . . PETIERVNT
The second diagonal either marks the end of the relative clause or the beginning
of the second half of the compound sentence. In paragraph 17 a space marks
the end of a series of three relative clauses, each depending on the word preceding it.
Parallel clauses of a compound sentence:
15. . . . IN · CONSVLÄTV · DECIMO... HS QVADRINGENOS. . .
PERNVMERaVI · /ET · CONSVL · VNDECIMVM · DVODECIM ·
FRVMENTÄTIÖNES · .. . EMENSVS · SVM · > ET ·
TRIBVNICIÄ · POTESTÄTE · DVODECIMVM ·
QVADRINGEN0S · NVMMÖS. . .DEDI ·
20. RIVOS · AQVARVM. . . R E F ß C l · ET · AQVAM · QVAE · MÄRCIA ·
APPELLÄTVR · DVPLICAVI ·
20. FORVM • IVLIVM · ET · BASILICAM. . .PERFECI · > ET ·
EANDEM · BASILICAM · CONSVMPTAM · INCENDIO.. .
iNCOHAVI · ET · SI · VIVVS · NON · PERFECISSEM · PERFICI ·
A B · H E R E D I B V S meis iusSl ·

24. STATVAE .. .QVAS · IPSE · SVSTVLI · EXQVE · EÄ • PECVNIÄ ·
DONA · .. .POSVI ·
A somewhat similar use is found in paragraph 16:
. . . sVmma sESTERTIVM · CIRCITER · SEXIENS · MILLIENS · FVIT ·
QVAM i>R0 · ITALIC IS · PRAEDIS · NVMERAVI · /ET · CIrCITER ·
BIS · MILLENS · ET · SESCENTIENS · QVOD · P R 0 · AGRIS ·
PRÖVINCIALIBVS · SOLVI ·
A principal use of punctuation is to set off two or more phrases that depend
on the same noun or verb from which one is separated by a subordinate clause
or other fairly long series of words. This is the kind of construction that is
sometimes explained grammatically as involving an "understood" repetition
of the noun or verb, e.g. the following could be explained as two parallel
clauses with a repetition of tradidi "understood" in the second:
27. EANDEM · GENTEM · . . .REGI · ARIOBARZANI · .. .REGENDAM·
TRADIDI · ET · POST · EIVS · MORTEM FILIO · EIVS ·
ARTAVASDI ·

46

THE 'EES GESTAE Α π β υ β Τ ΐ '

Similarly:
22. LVDOS · F E C I · Meo noMine QVATER · ALIORVM · AVTEM ·
MAGISTRÄTWM · VICEM · TER • ET · VICIENS ·
In paragraph 32, the punctuation (a blank space18 and two diagonal bars,
all clearly having the same function, with the fourth and fifth signs probably
lost in the lacuna) takes the place of a repetition of the word reges as an additional subject of the verb:
AD · Mfi · SVPPLICES · CONFVGenmi REGES · PARTHORVM ·
TIRIDAieS · ET · POSTea · PHRÄTes REGIS · PHRATIs FILIVs
MEDORVM · ARtavasdes adiabenorum ARTAXARES · /
BRITANNORVM · DVMNOBELLANVS · ET · TINcommius
sugambrORVM. • MAELO · /MARCAMAN0RVM · SVEBORVM
«egrimeRVS ·
Paragraph 33:
A · ME -GENTES · PARTH0RVM · ET · MEDÖRVm per legatos
PRINCIPES · EARVM · GENTIVM · RfiGfiS · PETiTÖS · ACCfiPERVNT ·'
PARthi vononem regis phrkTIS · FILIVM · RßGIS · OR0DIS · ΝΕΡ0ΤΕΜ ·
> MEDI · ARIOB Arzawem e. q. s.
In this classification we must place what is, of course, the most conspicuous
mark of punctuation in the entire inscription, which stands in the title to show
that the two genitive plurals are parallel and depend on exemplar:
RfiRVM · GESTÄRVM • D l V l · AVGUSTI · QVIBVS · ORBEM ·
T E R R A n m - IMPERIO · POPVLI · ROM
SVBlfiCIT · /ETINPENSARVM · QVAS · IN · REM · PVBLICAM ·
POPULVMQVE · ROmaNVM · FECIT · I N C I S A R V M 7
INDVABVS · AHENEIS · P I L I S · QVAE · SVrcT · ROMAE · POSITAE ·
EXEMPLAR · SBViECTVM ·
I t is possible that the punctuation at the end of the second line (not reported
by the editors, but clearly visible) was intended to show that the participle
modifies both of the nouns in the genitive plural, but of this we can scarcely
be certain without other examples of such use.
Comparable uses are:
15. PLEBEI · ROMÄNAE · VIRITIM · HS TRECENOS · NVMERAVI ·
EX · TESTAMENTO · PATRIS · ME I · Έ Τ · NOMINE · MEO · HS
QVADRINGENOS.. .DEDI ·
18

Or pos